Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
Researchers measure crowd-out around one level of charity output to identify whether giving is motivated by altruism and/or warm-glow. However, crowd-out depends on output, implying first that the power to reject pure altruism varies, and second that a single measurement of incomplete crowd-out can be rationalized by many different preferences. By instead measuring crowd-out at different output levels, we allow both for identification and for a novel and direct test of impure altruism. Using a new experimental design, we present the first empirical evidence that, consistent with impure altruism, crowd-out decreases with output.